| ricotz | darkxst, http://paste.debian.net/plain/225500 | 00:28 |
|---|---|---|
| darkxst | ricotz, so I tried to install 3.7 from testing on my Q laptop and hit the same 'focus issue' have been having with jhbuild on R :( | 10:07 |
| ricotz | darkxst, vanilla quantal with ricotz/testing ppa? | 10:54 |
| darkxst | ricotz, it was a vanilla quantal install, but has ubuntu-gnome-desktop installed | 10:59 |
| ricotz | with "focus issue" you mean the exclusive problem in g-s overview? | 11:00 |
| darkxst | yeh | 11:01 |
| darkxst | well not just overvew, happens with any pop-up or dialog | 11:01 |
| darkxst | i.e alt-tab | 11:01 |
| ricotz | so switching to a window with alt-tab shows it but doesn't focus it? | 11:04 |
| darkxst | mabye focus is really the wrong word, but basically launching a dialog (say alt-tab) all input gets stuck keyboard/mouse, that times out after about 10-30sec | 11:05 |
| darkxst | then you get one more input event from key/mouse and it gets stuck indefinately | 11:05 |
| ricotz | ok, focus-issue is the wrong word for it then | 11:06 |
| ricotz | i might have seen this with the volume control with raring | 11:07 |
| darkxst | also its not your typical push/pop modal issue | 11:07 |
| ricotz | the alt-tab window stays there for ever? | 11:07 |
| darkxst | that tends to cause wierd things with events going to the wrong windows | 11:07 |
| darkxst | ricotz, yeh | 11:08 |
| ricotz | could you use your volume conrol keys in a "faster way" | 11:08 |
| ricotz | so tune it up and down a bit | 11:08 |
| darkxst | I did not try volume keys | 11:09 |
| ricotz | this is how i am seeing it occasionally and the volume-control-window of g-s gets stuck | 11:09 |
| ricotz | of course this window doesnt use modal so it doesnt grab all keys afterwards | 11:09 |
| ricotz | but it sounds like the same issue | 11:10 |
| darkxst | similar, except this is affecting basically everything | 11:10 |
| ricotz | i guess you told me already, which x-driver do you use? | 11:10 |
| darkxst | intel on Q, nvidia blob on R | 11:11 |
| ricotz | ok | 11:11 |
| ricotz | blob and intel on R here | 11:12 |
| ricotz | only seen my issue on with blob | 11:12 |
| ricotz | will try to reproduce it (i updated today to 313.18) | 11:13 |
| ricotz | also i am using a pre-release of eglibc 2.17 | 11:13 |
| ricotz | on raring | 11:13 |
| ricotz | which solves some deadlocks of 2.16 | 11:14 |
| darkxst | of course the strange thing is the 3.7 g-s packages from testing work perfectly on my R machine with broken jhbuild | 11:18 |
| darkxst | and when I attached to g-s with gdb, it just had g-s sitting in a normal idle state | 11:21 |
| darkxst | anyway I am stuck on my laptop this week, so can't really go and break it again, to try debug more | 11:24 |
| ricotz | darkxst, alright, this seems to be caused by some dependency or the X stack | 11:46 |
| darkxst | hm right | 11:59 |
| darkxst | btw, I think I probably have xorg-edgers installed on my R/jhbuild machine | 12:14 |
| ricotz | me too ;) | 12:40 |
| jbicha_ | gdm unlock seems broken http://paste.ubuntu.com/1542043/ | 17:12 |
| darkxst | jbicha_, I think that is this bug https://bugzilla.gnome.org/show_bug.cgi?id=689106 | 20:39 |
| ubot5 | Gnome bug 689106 in lock-screen "ScreenShield: try harder to become modal, and catch failures" [Critical,New] | 20:39 |
| jbicha_ | darkxst: thanks, I guess I should file a bug report then | 20:40 |
| jbicha_ | I'm getting the gnome-screensaver unlock and the gnome-shell unlock but then it gets stuck with just the gray gnome-shell unlock screen (I'm using 3.7.4) | 20:41 |
| darkxst | jbicha_, how come you are getting the gnome-screensaver unlock? | 20:48 |
| darkxst | the 'incorrect pop' results in basically a stuck grab? | 20:53 |
| jbicha_ | uh I think so | 20:53 |
| jbicha_ | let me reboot just to clear some things | 20:54 |
| jbicha | darkxst: ok, never mind it seems to be working fine now | 20:59 |
| jbicha | it could just have been me restarting gnome-shell manually which confuses the too fragile gdm/gnome-shell combination | 21:01 |
| darkxst | oh, its possible that is gnome-screensaver is running when you run 'gnome-shell --replace', then you end up with both | 21:05 |
| darkxst | I have also occasionally seen g-s fail to get the auth channel to gdm, but pretty sure that results in g-s lock being disabled | 21:06 |
| darkxst | speaking of stuck grubs I just got hit by one, but different | 21:13 |
| darkxst | possibly caused by a notification, while in (or maybe exiting) overview | 21:15 |
| ricotz | jbicha, hi | 21:35 |
| ricotz | darkxst, hi, was the hg patch usable? | 21:35 |
| darkxst | yeh | 21:36 |
| ricotz | jbicha, i have taken a look at g-c-c and it would be quite some work to get a external library again and make g-c-c aware of the ubuntu plugins | 21:37 |
| ricotz | jbicha, the easiest would be to consume the real plugins as a patch and integrate them like the upstream ones | 21:37 |
| jbicha | ok I got the login screen problem again, I just had to wait for the screen to automatically lock | 21:47 |
| jbicha | ricotz: yes, g-c-c is quite a bit more difficult to get the external library working again | 21:48 |
| jbicha | I'll let Canonical fix that if it's a priority to them | 21:52 |
| ricotz | btw, i pushed empathy, but i guess i will split account-plugin-* again with the next release | 21:52 |
| ricotz | jbicha, currently the online-accounts are not usable which is a big deal :\ | 21:53 |
| ricotz | jbicha, or got they integrated in g-c-c? | 21:55 |
| darkxst | jbicha, yeh the failed push happens when the idle timer kicks in | 21:55 |
| jbicha | ricotz: ah, well for the staging PPA we may want to disable uoa then since I doubt g-c-c will get fixed until Canonical cares, which probably won't be for a few more months | 21:55 |
| jbicha | darkxst: should I open a gnome bug for that? | 21:55 |
| darkxst | jbicha, it is the same bug I linked you before | 21:55 |
| ricotz | jbicha, uoa is upstreamed | 21:56 |
| ricotz | meaning in empathy | 21:56 |
| darkxst | although those patches really just workaround the issue and don't fix whatever is causing the grab to fail | 21:56 |
| jbicha | ricotz: right but it's still a configure flag | 21:57 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!